Visual Electric: finally a real graphical interface for AI image generation


Visual Electric is an online tool that generates images using artificial intelligence. The latter works on the same principle as Midjourney, namely entering a prompt in the form of keywords describing the content of the illustration to be created. On the other hand, where specific jargon has been developed for Midjourney, Visual Electric has a real graphical interface allowing the user to specify their parameters.

Consistent with query entry

Visual Electric notably uses generative AI to develop prompt suggestions. As the user enters their query, the AI ​​emits suggestions consistent with the context. For example, by providing “a medieval castle with ducks as guardians”Visual Electric offers several ideas to complete the request, such as “bathed in an atmosphere of moonlight”.

Prompt suggestions by Visual Electric.

© Visual Electric

Generating an image with visual electric

Generating an image with Visual Electric.

© Visual Electric

A button allows you to select the desired style of the generated image, each atmosphere having a rendering example. This is followed by the choice of the ratio, again using buttons integrated into the Visual Electric graphical interface.

Visual Electric edit an image

Edit an AI-generated image with Visual Electric.

In our example, we thought that the prompt had not been formulated in a way that the AI ​​could understand. We actually wanted to get a medieval castle with ducks as guardians. Visual Electric allows you to modify this prompt via the Remix functionality. Our prompt now looks like this: “A medieval castle with enormous ducks as warrior guardians, bathed in an ambiance of moonlight”.

Unfortunately, despite the modification of the prompt, Visual Electric’s artificial intelligence still seems not to understand our request.

Image generated with Visual Electric AI

Image generated with Visual Electric AI

© Visual Electric

We wanted to check the understanding of the same prompt by Dall-E, the OpenAI AI integrated into ChatGPT. For this example, Dall-E performs much better, as demonstrated in the illustration below.

Image generated with Dall-E AI

Image generated with Dall-E AI.

© Dall-E / OpenAI

If Dall-E seems more efficient in its understanding than Visual Electric, the latter tool has a much more promising interface. In the same vein, Dall-E is only accessible by paying a subscription to ChatGPT, where Visual Electric offers a free version accessible to all.

Advertising, your content continues below



Source link -98